Estrogen receptor positive (ER+), or luminal type, tumors account for 60-75% of all breast cancers and for more deaths than all other types of breast cancer combined. Aromatase inhibitors (AI) are recommended as first-line adjuvant therapy for hormone responsive tumors in postmenopausal women. In spite of significant success with hormonal therapies, ~20-25% of patients with ER+ disease will progress by 10 years with relapsed patients ultimately succumbing to their disease. Poor drug adherence, principally due to intolerance to side effects, remains a major challenge for achieving greatest drug benefit. A need to maximize AI efficacy is evident. Non-steroidal anti-inflammatory agents (NSAIDs), particularly sulindac, demonstrate potent and mechanistically supported anti-cancer activity for breast tumors in preclinical models. We hypothesize that sulindac, combined with AIs may act synergistically on breast density and breast tissue biomarkers as surrogates for relapse risk. The addition of sulindac to AI therapy may also have the added benefit of decreasing muscle and skeletal pain associated with AI use and thus improved adherence and long-term efficacy. To test our hypotheses, 150 breast cancer patients, stable on AI therapy for ER+ tumors, will be randomized to one of two intervention arms for 12 months: 1) AI + sulindac 150 mg bid or 2) AI + placebo bid. Our specific aims are: 1. To compare change in breast density as measured by Magnetic Resonance Imaging (MRI)-acquired fat-to-water ratio (FWR) (primary trial endpoint) within individuals and between treatment arms. We hypothesize that women treated with AI + sulindac 150 mg bid will show decreased breast density (i.e., increased FWR) over 12 months, whereas breast density in women receiving AI + placebo will not change. 2. To compare the apparent diffusion coefficient (ADC) of water within individuals and between treatment arms. We hypothesize that ADC values measured by diffusion weighted MRI (DW-MRI) will significantly change in women treated with AI + sulindac 150 mg bid over 12 months, whereas they will not change in women receiving AI + placebo. 3. To compare pain scores using the Brief Pain Inventory-Short form (BPI-SF) within individuals and between treatment arms. We hypothesize that women treated with AI + sulindac 150 mg bid will experience reduced pain scores over 12 months, whereas they will not change in women receiving AI + placebo. In addition, because the prodrug sulindac sulfoxide (Clinoril) has been shown to spare renal synthesis of the vasodilatory prostaglandins in patients with normal renal function, we hypothesize that daily sulindac use will not increase blood pressure (BP) in women on AIs with normal renal clearance and thus, will not elevate risk of CV toxicity mediated through drug-induced hypertension. Success in this phase II biomarker trial of sulindac combined with AI will serve as justification for a larger trial with cancer specific outcomes.